Chapter 9 Google's formula for success

The Way of Google's Success——New Innovation Leading to "Network Survival" introduction In high-tech enterprises, everyone is talking about innovation, and everyone hopes to use innovation to create better products and obtain more economic returns.However, there are not many companies that can truly combine innovation and product development and benefit from it, and there are only a handful of truly efficient innovations that can profoundly change technological development trends and even change people's lives.Innovation is indeed a very challenging thing, and it is also very worthy of people's deep thinking.

Taking my own experience as an example, when I was researching speech recognition technology at Carnegie Mellon University, I didn't think too much about how to turn the technology into a product that could be used by everyone. Like many researchers, they believe that if new methods or technologies can be discovered and more papers published around these methods and technologies, many companies will definitely pay attention to them and apply them to their own products.Although the research I did represented the highest level in the field of speech recognition at that time, and my thesis was unanimously recognized by the academic community, I did not see my technology transforming into actual products at that time.On the one hand, this may be because my thesis is highly academic, and it is not easy for ordinary technicians to understand. Applicability and optimization methods of these technologies for specific application areas.This kind of personal experience made me understand that those purely academic innovations that are divorced from industrial practice are very difficult to succeed.

After arriving at Apple, the company attached great importance to my speech recognition technology, and specially arranged a technical department to be responsible for developing speech recognition products.At that time, marketers and product developers were very eager to bring speech recognition products to the market, but they did not consider that this technology needed a necessary development and stabilization stage from birth to maturity.Due to the lack of sufficient research and development accumulation, it is difficult for products launched prematurely to achieve due success in the market.

Later, when I was working at Microsoft, I found out that Microsoft created a different management system for researchers and developers. Researchers are engaged in future-oriented innovation in Microsoft Research, and developers are in the product department to develop software products for the market.The research results of researchers can be transferred to the product department through the interdepartmental cooperation mechanism.This method of parallel management according to the characteristics of different jobs can better solve the problem of disjointed research and development, and can also stimulate the enthusiasm of research and development personnel.However, because the goals of the research institute and the product department are different, the methods of measuring performance are inconsistent, and there are considerable differences in the thinking and working methods of researchers and developers. In the actual process of R&D and product conversion, there will still be many frictions , Procrastination and other things that affect efficiency.

The R&D management systems of Apple and Microsoft are very different. The efficiency of the two may not be optimal, but they are accumulated during the long-term practice and development of their respective companies, and they are the most suitable for the actual situation of the company. The R&D system needed for development.For example, Microsoft's R&D parallel system is one of the best management methods I have seen for software R&D. However, in the emerging and ever-changing Internet industry, we need a more efficient R&D management system that is closer to the needs of Internet users. Google has achieved great success in the Internet market, which is inseparable from its unique research and development mechanism.On the one hand, the Internet is a brand new world, and related technologies and products are changing every day, which requires related companies to closely combine technological innovation and product development; A platform for free communication, which also provides a new way of thinking for innovation: we can rely on everyone's strength, encourage collective participation in the research and development process, and quickly collect and respond to user feedback information through the Internet after the product is released. Google has taken these two issues into consideration and created an open, equal, free, efficient, integrated R&D, user participation, and rapid response R&D management system.

This paper discusses Google's innovation and R&D mechanism facing the Internet age in a detailed and in-depth manner, as well as Google's unique corporate culture and talent strategy.It should be noted that I do not want all Internet companies to imitate Google's system—any system has its advantages and disadvantages, and only the most suitable management method that can best meet actual needs can achieve the greatest success. There is no doubt that innovation is one of the most passionate and dynamic words in today's era.Looking at innovation alone, people may regard related things as the sum of extraordinary work done by intellectual elites driven by curiosity or utilitarianism, but if innovation is placed in the industrial environment, It is not difficult to find that innovation is actually a step-like progressive form that integrates science and technology, the self-renewal law of social economy and historical and cultural factors. The "three-dimensional" phenomenon with deeper connotations such as internal mapping in the new economy.

Although everyone advocates and encourages innovation, few people have deeply analyzed the regular content behind this "three-dimensional" phenomenon.For example, different types of enterprises will always choose different innovation models. Are there any rules to follow?If innovation itself is closely related to industry, era and economic operation, then in the latest, coolest and most dazzling network economy, what kind of innovation can create the greatest value?For a new technology company that regards innovation as its soul, how to give full play to the value of cultural factors so as to stimulate the innovative spirit of the company more effectively and directly?

With the gradual development of the IT industry, according to the needs of different periods, several popular innovation models have gradually formed in the industry.When the Internet age comes today, an "innovative concept" or "innovative culture" represented by Google that best matches the spiritual connotation of the Internet is gradually emerging, and it has created miracles one after another.For Google, these miracles include a huge user base (1 billion page views per day); unprecedented growth rate (more than 100% annual growth rate, it took only 7 years to make the company's market value reach 100 billion US dollars ); high employee satisfaction (close to 0% employee turnover); efficient product production mechanism (since the company was founded in 1998, it has launched more than 50 successful products)... The most incredible thing is that "Oxford English Dictionary (OED) and Merriam-Webster's Dictionary (Merriam-Webster) have included "Google" as an official word, which shows that unlike previous IT companies, Google uses the platform of the Internet to not only It provides users with a product to use and creates a mass culture.

With the help of my rich experience in R&D, operation and management accumulated in the IT industry, the author will discuss some of the most popular innovation models in the IT industry in a simple and simple way, and focus on introducing to you Google, which is at the forefront of the Internet age. Innovation culture", and the "core wealth" in this new model: talent and Google's "talent concept". Innovation model and "zeitgeist" As an important part of human society and historical activities, the innovation of science and technology inevitably has a strong "time color".Take the development of physics as an example. When the young Newton returned to his hometown in the countryside in 1666 and independently completed several pioneering research results, he would never have imagined that more than two hundred years later, Maxwell, JJ Thomson Or the scientific research of Rutherford and others would be so strongly dependent on the process of the industrial revolution in the same period, and he would never have imagined that after more than three hundred years, the working methods of Gell-Mann, Hawking, Yang Zhenning and others would be different from those of others. Large-scale international collaboration and the rapid development of computer and network technology are closely related...

Applying this "color of the times" to the IT industry, we can probably assert that: Technological innovation promotes the industrial process, and the industrial process determines the business model of the enterprise, and the business model of the enterprise in turn further affects or determines the innovative ideas in the same period and in the same field. For example, before the arrival of the wave of the Internet, there were four mainstream innovation models in the IT industry - "ivory tower" innovation, "one-way output" innovation, "brain rent" innovation and "R&D interactive" innovation.They have created brilliant scientific and technological achievements and corporate models in different historical periods. Although they cannot better adapt to the new challenges of the Internet age, they still play an important role in their familiar fields until today.

"Ivory Tower" Innovation In the early stage of IT technology development (before 1980), the relationship between technology itself and the industrial chain was not particularly close.Software and hardware R&D personnel are more used to simply considering the scientific and technological content of R&D targets, and do not pay attention to or temporarily ignore their industrial value.This "ivory tower" innovation model can also be called "head-slapping" innovation in a certain sense-the R&D personnel have inspiration, write papers, publish them in core journals, and then get recognition from peers . Obviously, this innovation model can give full play to the creativity and imagination of R&D personnel, but at the same time, it has the disadvantages of being far away from the needs of the industry and making it difficult to transform R&D results into actual products. "One-way output" innovation There is an American movie that tells such a story: a person suddenly dreams of building a baseball stadium, and at the same time he firmly believes that after the stadium is built, people must flock to it, "Build it and they will come!" Of course, the final result Maybe not as good as he imagined. This practice of inventing first and then attracting users was very common in the IT industry from the 1980s to the 1990s.Many well-known companies or research institutes have achieved remarkable success by relying on this innovative model—initially leading the product design process by scientists or technicians, and then putting effort into bringing the product to the market—but many others have adopted similar approaches. of businesses falter either because they misunderstand potential user needs or because of poor marketing.Today, fewer and fewer companies adopt this innovative model. "Brain for rent" innovation When the .COM bubble burst, many companies realized that in order to maximize the satisfaction of user needs with the minimum cost, they must change according to the market and customers.They have transformed their R&D work into market-oriented or customer-oriented "R&D services": the needs of customers are directly reflected in the results of market research, and the results of market research determine what products and technologies the company will develop. Contribute your own "brain work" around this principle-so, this innovation model is also called "brain-rental" innovation. The advantages of this innovative model are fast response, highly targeted product design, and high conversion efficiency of R&D results.But on the other hand, because R&D personnel are almost directly facing the pressure from the market and customers, in their work, there are more and more service and customization content, and the part that really belongs to "innovation" must be less and less. "R&D interactive" innovation In order to solve the contradiction between research and development, some companies that have grown rapidly in the PC wave have set up research institutes or similar research institutions outside the product department.The research institute has an independent structure and does not need to work under the pressure of the market or product department. Its mission is to create the latest technology; the product department is responsible for responding to market demand and transforming the results obtained by the research institute into excellent product.At the same time, the company's top management will coordinate the relationship between the research institute and the product department to achieve the best "interaction" between the two. The innovation of "Interactive R&D" has freed many companies from the problem of disconnected R&D and R&D, while maintaining the speed of response to market demand and unremitting exploration of future technologies.Of course, this innovation model also has its problems. For example, the cost of communication between the research institute and the product department is high, the model for resolving conflicts and increasing understanding is not clear, and the company’s management investment in the two is relatively large. New Challenges in the Network Era In recent years, after getting rid of the Internet bubble, the IT industry and the Internet have ushered in an unprecedented development opportunity: some business models that have been in the exploration stage for a long time have begun to make profits and develop rapidly, and new market demands and new user groups are constantly emerging. The emergence of new technologies and new products has greatly shortened the listing and life cycle, the capital market is increasingly dependent on technological innovation, and the competition and change of enterprises are getting faster and faster... In this real Internet age, the network Enterprises with services and network products as the main direction need a brand-new innovation model. Only in this way can enterprises maintain the greatest flexibility and the highest work efficiency in the face of new challenges. New Innovations Leading to "Cyber ​​Survival" As one of the most successful network service and network product providers, Google has embarked on a road of innovation that is completely different from the traditional model since the day it was founded.It can be said that Google's innovation model has important typical significance in today's IT industry: it not only successfully solves the problem of disconnection between research and development, but also achieves the optimal management cost and R&D efficiency. The success achieved on the Internet is the best testimony of the value of this innovative model. R&D integration As early as the 1930s, the German philosopher Heidegger predicted when discussing the relationship between science and truth: "Researchers must spontaneously flock to the scope of the essential form of technicians in the fundamental sense. Only in this way can he maintain Active ability, so that it can truly exist in the sense of its time and not be outdated." (Martin Heidegger, "The Age of the World Image", 1938) Objectively speaking, the boundary between research and development has been in the last 50 years. It is quite blurred here, and this kind of blurring—more precisely, "integration"—is the inherent need of industrial development. If you have to separate new technology research and product development, you may encounter some inevitable problems (mainly reflected in the communication and unified management between the two departments), and these problems have gradually emerged in companies that adopt the separation of R&D come out.Due to the different division of labor, people in the research department mainly focus on some pioneering and revolutionary projects. When they get meaningful ideas, they may not be recognized by colleagues in the product department until a certain day , when the “integrated R&D” competitor made such a product, the product department realized that it had missed the opportunity.And sometimes, in order to provide better services to users, when the product department hopes that the research department will do some things that users need, but the research department disdains to do this kind of "unchallenging" work and loses more. opportunities for users.What's more, there is still a feeling of mutual contempt between the two departments, because people in the product department think they are "the main force making money for the company" and despise colleagues in the research department, and people in the research department think they are "the main force for making money for the company". "Geniuses with high education" are superior to others, which also creates obstacles for communication between the two departments. But in Google, there is no separate research department and product department, and all engineers wear two hats, R (research) and D (development). Google requires every engineer not only to have the courage and intelligence to innovate continuously, but also to have the skills and experience to turn their ideas into reality.In this paradise of innovation and practice, anyone can come up with a brilliant idea at any time, and everyone has the opportunity (or obligation) to turn their ideas into reality.This approach of integrating research and development completely eliminates the gap between innovation and practice, and at the same time saves management and communication costs to the greatest extent, and improves work efficiency. Doing it yourself, that is research and practice, can sometimes give people more inspiration, get completely different ideas from the past, and make some technical problems easier to implement. The most convincing example is the image search proposed by Google engineer Zhu Huican s solution.In previous pure academic research, image search was a difficult problem. People often used color difference and advanced signal processing techniques to find a picture that best matches the query. Although these techniques are "cool", they are not yet mature.In the past ten years, many world-class research institutions have published many good articles in this field, but because most of these researchers are out of touch with product development, these research results have not been successfully translated into products.Combining his own expertise in parallel computing and his partners' expertise in information retrieval, Zhu Huican proposed that "on a web page, each picture has a title and surrounding text", and that "the text linked to the picture (anchor text) generally affects the picture. There are a lot of clues in the content of the content”, these titles, text and clues are very easy to extract, using this idea, quickly made “the best image search tool on the Internet”-Google Image Search.This example shows that a team can quickly produce good work under the premise of combining practice.Of course, those advanced technologies are not completely useless, but they need to be combined with specific environments, such as image retrieval using signal processing technology. In order to achieve practical results, the amount of search data should not be too large. Recently, Google began to target personal It is possible to use this technology to achieve "face recognition" in the photo album. The founders of Google, Sergey Brin and Larry Page, once had a conversation to discuss why Stanford’s Ph.D.s generally perform better than postdocs. The final conclusion is that Ph.D.s can choose topics they are passionate about. Follow your favorite teacher to do projects, but postdoctoral fellows can't choose.Similarly, in a company, if "R&D is integrated", then engineers can follow up the project from conception, to research, to implementation, from beginning to end, which will enable them to have a sense of ownership of the project, and thus be more enthusiastic and care about the project. User experience, make the project better. The only disadvantage of this "integrated R&D" model is that it is not suitable for talents who can only innovate but cannot practice, or are strong in practice but weak in innovation.There are many good scientists in this world who are not engineers, and there are also many good engineers who are not good scientists. These people must go through the necessary learning and training to adapt to an environment like Google. individual freedom Many people know that Google provides each employee with 20% of their own time.But people may not know that this 20% of the time is actually a crucial part of Google's innovation model.Once there is this 20% disposable time, the ideas hidden in the minds of engineers will emerge in an endless stream. Under the guidance of creativity and imagination, the value of engineers can be reflected to the greatest extent—many things that make Google Proud products like Gmail and Google News are created 20% of the time by engineers. In this 20% free time, employees can do anything, including study, training and even writing books, but the freedom of employees is not limited to this 20% time.Although employees need to spend 80% of their time on the tasks assigned by the company, these project companies generally only decide on one direction, while the specific implementation plan and technology selection are determined by individuals, so this 80% of the time also reflects personal freedom.Moreover, for a specific project, whether it takes up 80% of the individual's time or 20% of the time, this ratio is not static.When a future project is just a personal idea, it may only take up 10%-20% of the working time of the person who proposed the idea to realize it; After more people join in, this project may take 30% to 20% of the time of the participants until the product prototype is made; the product prototype will be placed in the "big laboratory" of the Internet, and it will be tested and tested by the majority of users. After getting support, this project may become the main work that project team members need to complete 80% of the time, and this state will continue until the birth of new products; when new products are launched, new ideas appear , and repeat such a process (as shown in the figure below). In addition, Google does not try to keep engineers fixed in a project or a product group for a long time like ordinary enterprises.Engineers can easily switch to the group work they are interested in, and can also join the development process of several products at the same time.Obviously, this kind of flexibility in management can also better stimulate everyone's sense of innovation. The reason why Google adopts this management style is because the company recognizes that the best talents do not want to be completely controlled by others or the company, but want to have the opportunity to control their own destiny.The direction of the company's progress can be managed macroscopically by clarifying the company's mission ("integrating global information and making it accessible and beneficial to everyone") and reasonable project allocation (70%-20%-10%), but in How to do the actual project, determine the product function or project timetable, or propose new ideas, etc., you might as well give employees more power to choose. equal participation Equal participation in the research and development process is another cornerstone of Google's continuous product innovation. Google has established an "idea library" with a rating and ranking mechanism within the company.Employees can put their product or technical ideas into the "idea library", and then voted by other employees.The company selects future research and development directions from the "idea leaderboard", and determines the scale and importance of the project according to the popularity of each "idea".For each specific project, engineers can not only vote with their hands in the "idea library", but also vote directly with their "feet", that is, they can personally participate in the research and development of the project, or participate in the company's internal testing , Realize your technical dreams with your colleagues.When the product prototype is realized, it will be placed in a Google "lab" and made available to anyone on the web.The result of this is that after almost every product of Google comes out, it will be tested by hundreds of millions of users all over the world, and these users are "netizens" distributed all over the world. Google believes that when a product is a "good" product that is useful to everyone, most users will be very happy to use the product, then this product will have a high click-through rate or download rate; and vice versa.For good products, the company will further develop and launch the Beta version until the official version is released. Google deeply understands that this kind of "voting" from end users with the mouse is the most equal and most effective "optimization method" to guide the R&D process. (As shown below) Real-time and efficient In the past, product research and development always went through many processes such as market research, product design, development and testing, product packaging, and marketing promotion. It is quite difficult to do market research, and it takes a long time. It is very likely that when you start to make products after getting the survey results, you will be far behind by your competitors. Google takes advantage of the real-time characteristics of the Internet and regards the entire Internet as its own laboratory.When engineers have a new idea, they immediately implement the initial version of the product, and then put it in Google Labs to accept users' trials; while the first batch of users are trying the product, the frequency of users' downloads and usage can increase. Help the company judge which product is the user's favorite; the user's feelings and suggestions for the product have been continuously fed back to the engineer through the network; the engineer can always use the information obtained in real time to modify or upgrade the product, so that the product in the laboratory matures as soon as possible and Google will decide which products can "graduate" from the laboratory based on user feedback.If you can't decide whether a product should have a certain function, according to the past practice, it often requires special manpower to do market research or interview target users, and it will take at least a month or more to reach a conclusion.At Google, those talented engineers only took two or three days to make the new function first, and provided this function on half of the servers, while the other half did not provide it, and finally decided whether to include it based on the user's trial and feedback results. This function, which greatly shortens the time it takes to make a decision, and the conclusion is very accurate. At Google, we may have many projects going on at the same time, and many products on the web, but not because the company has enough money to waste.In fact, many of these projects are just an attempt, not necessarily the company's future development goals. The reason why they put them online when they are uncertain is because the company realizes that the cycle of getting user feedback in this way is shorter than doing a market survey. Or target user access is much faster.Instead of doing market research blindly, but using existing network resources, combining market and technology, and getting market feedback in real time, this is very consistent with market rules. user priority Many software companies follow the principle of user first, but the word "user" has an additional meaning in Google.Some people criticize Google for being technology-led rather than user-first. This is a misunderstanding, because Google's products are aimed at Internet users, and Google employees are actually Internet users. That is to say, engineers are developing products while , and can also examine and evaluate products as end users, so doing it by them is user-led. All of Google's products are launched on the Internet, get feedback from users in real time, get statistics on users' downloads and use of products in real time, and modify or upgrade the company's services or software in real time based on these feedback and data. In this way, the R&D process of most of Google's products will have the direct and full participation of end users: in the product design stage, the needs of end users can be reflected to the R&D team without any obstacles; Software products are constantly tested by end users; during the product stabilization and release phase, engineers from within the company and users from all over the world jointly form an efficient testing team.Products produced through the process of "engineer innovation, user voting" will naturally achieve excellent results in user satisfaction. Although Google has adopted an engineer-led research and development mechanism, its product innovations have utilized the characteristics of the Internet to involve end users from beginning to end, so it is a "user-centered" innovation. Google's adherence to the principle of "user first" is also reflected in the projects implemented by the company. Google develops products based on "whether this product is useful to users", not "whether this product can make the company money".For example, Google's vigorously promoted book search project, which costs a lot of money.And because it's free to the end user, there's no chance the company will be rewarded in the short term.Other companies that are primarily for-profit might put people on more rewarding projects, or spend money on hiring more salespeople instead of engineers.But Google believes that this project can help users find any book, no matter whether the book has an electronic version before, whether it is on the Internet, and does not need to know the title or author name, just use any word in the book to search Yes, this is what users want, so we will continue to do it. clear mission Microsoft's corporate mission thirty years ago was "to put a PC on every desk", so it merged four successful products into Office suite products more than ten years ago to promote the realization of this mission, and finally Gained a high market share; when Ford designed a new car, in order to achieve the corporate mission it set at the beginning of its establishment - "let every ordinary American afford a car to enjoy "Beautiful life", abandoning the design that emphasizes the performance of the car, adopting another design that emphasizes the cheapness and comfort of the car, and finally became a giant in the automotive industry.Like these most successful companies, Google also has a clear and clear corporate mission, which is to "integrate global information and make everyone accessible and benefit from it", and use this mission to guide resource allocation and guide the development direction of the company . Some people see many of Google's new products, such as Google Video, GMail, Google Desktop Search, etc., and don't understand the inner connection in these products, so they wonder if our product line is not messy, and engineers can do whatever they want what.In fact, this is a misunderstanding, because all these products are in line with Google's mission.Because facing the astonishing amount of information content on the Internet that obviously lacks order, ordinary Internet users will always feel at a loss.How to present the most useful information in the most appropriate form to the users who need it the most should be the biggest challenge in the Internet age. Seeing the huge opportunity hidden behind this challenge, Google spared no effort to invest its research and development efforts in information integration and network services. Google's goal is to provide a unified and convenient organization and retrieval channel for different forms of information such as text, pictures, news, emails, commodities, music, and videos distributed all over the world. This goal is exactly what Google is planning and The most important decision-making basis when formulating the direction of innovation. (The picture below shows the family of products that Google has launched to users that meet Google's mission) A company has a clear mission. In addition to being of great help to the company itself when making important decisions, it can also motivate employees and enable everyone in the company to better understand the meaning of their work. Let culture become the "source power" of innovation It can be said that Google's innovative model has created a new development path for the IT industry in the Internet age.But what can we discover if we go further to find the driving force behind this innovation? The culture and values ​​of an enterprise are its spirit, way of thinking and behavior, and a code of conduct and values ​​formed by all members of the enterprise in the process of production and operation. Google's sensitivity to innovation and its desire to establish a new generation of innovative models are fundamentally derived from Google's unique corporate culture.In fact, the corporate culture of each company will more or less affect the behavior of the company, but in Google, many corporate cultures themselves have a strong "innovative" color. equality, mutual respect The company empowers each employee to take charge of their own work.The company does not have a "punching card" system, and everyone's commute time is basically determined by himself. At Google, all employees are equal, senior personnel basically have no "privileges", and project directors and technical managers still have to personally participate in specific technical work.Within the company, anyone can talk to anyone about any topic and email anyone.An egalitarian company reduces information congestion within the company and increases ownership by all employees.For example, my office door is always open (unless there is noise that disturbs others, or that disturbs me).Anyone with anything can come to me directly. Once, a secretary who had just joined Google didn't know how to use the fax machine in the company, so he asked a gentleman who was waiting to send a fax behind him to teach him how to operate it.After successfully sending the fax, the new secretary thanked the enthusiastic person, and the two reported their names respectively.To the surprise of Miss Secretary, the enthusiastic person turned out to be Schmidt, the CEO of Google. When CEO Schmidt first joined Google, all Google employees did not have their own independent office, but the employees still felt that it was necessary to give him a relatively quiet office space, so they arranged for him a very small independent office.One day, an engineer came to Schmidt's office by himself and said, "Everyone else is in a shared office, and my side is too crowded, so I want to come here from you." Schmidt was surprised and asked him: " Did you ask your boss?" The employee went to ask the boss and came back and said, "The boss thinks I should sit here too." So they shared an office until the company bought a bigger building .To this day, apart from a few dozen employees who "have meetings with people all day," all employees still share offices. At Google, everyone has to deal with their own things, even the CEO is no exception.It is precisely because of such a corporate culture that Google's R&D personnel have naturally formed an innovative concept of integrating R&D. 我们的很多决策都是由工程师集体决定的。有一次我们开会讨论一个应聘者，他是非常优秀的学生，但是成绩很糟糕，有几科只得60分。虽然我很想雇用他，但是有一位工程师提出严重的顾虑：“成绩不是最好的没有关系，但是这么差的成绩就是一种与态度或责任心相关的问题了。”当时，有半数的工程师基本上同意他的看法。在很多其它公司，老板在这种情况下可能不会采纳该工程师的意见，但是在Google，我决定遵循公司平等的文化。当时，我问这位工程师：“那么，我们可不可以把决定权交到你手上，你打一个电话给这位应聘者，如果他不能说服你他是一个负责任的人，我们就不要雇他？” 创新、实干 著名的Google News其实是源自一位工程师的“灵机一动”：他希望看到最热门的新闻，但又不满意门户网站编辑的筛选结果，于是就亲自动手，实现了一个由软件自动编选、排序的“新闻中心”。这种创新加上实干的企业文化直接影响了Google公司对新产品和新技术的思考方式。 其实，在Google，创新文化已经深入到了员工身边的每一件事。有一次，公司的一个老板告诉自己手下的一个团队说，你们只要达到目标，我就送你们一个游泳池。当时大家都觉得这不可能，因为在高楼大厦中间没有足够的地方建游泳池。结果，那个团队真的达到了目标，这个老板第二天就把游泳池带来了——不过这是一个吹气的游泳池。后来，Google真的建了一个游泳池，而且很创新的，只有四米长。四米长怎么游泳呢？它有一个逆流喷水装置，让你永远在同一个地方游，跟跑步机的原理一样。这样的游泳池也需要一个救生员——不是怕游泳的人溺水，而是怕可能有撞到头的危险。 Google的两位创始人也不会忘记发明一些“新奇”的玩意儿来为工作增添些乐趣。有一次，拉里·佩奇自己动手，将装有自己开发的测试程序的笔记本电脑安装在可以遥控的玩具车上，然后蹲在地上，指挥着自己的测试车跑遍公司的各个角落——其目的竟然是为了测试公司内部的无线网性能。 创新引导实践，实践支持创新。实践和创新缺一不可，这就好比只懂得力学原理的人和只知道铺砖叠瓦的人都无法独立建起一座摩天大厦一样。同样的，在新的世纪里，也只有那些善于将创新和实践结合起来的人才有可能获得最大的成功。 透明、客观 在Google，所有员工，所有研发小组的工作进程、工作业绩都会在内部网上公布，每个人都可以看到其他人在做什么，做得怎么样。考核员工业绩时，Google并不仅仅依赖管理者的评价，反而更看重与该员工一同工作的其他人的意见。透明和客观的管理方式可以营造出公平的竞争环境，并进而最大限度地激励员工的创新热情。公司的任何员工都可以发“奖金”给别的员工，来感谢他对你的项目的帮助。 在公司内部，没有什么秘密的项目。所有的项目都放在内部项目管理系统中，让所有其它的人能够看到。公司开会任何人都可以来。有一次郭去疾去开会，结果十分钟后发现来的都是副总裁，还有CEO和两位创始人，这时他才意识到自己走错房间了。但当时没有人问他是谁，也没有人赶他走。 公司每个星期五都有TGIF活动，在该活动中，管理者把最新的方向、信息、产品、决策等及时地介绍给所有员工。在总部，这个活动由两位创始人亲自主持，在中国就是我和周韶宁亲自主持。 许多公司便大了就会有所谓的“办公室政治”出现。比如，一个团队自己做得很好，但是却会和别的团队发生冲突，或者是一个领导在背后批评另一个领导。我们如何解决这个问题呢？我们要求每一个技术总监不要和自己部门共享办公室，而要和与他合作或工作重叠得最多的另一个技术总监分享办公室。这样一方面两个人可以保持合作，另一方面，也不提供机会让人可以关着门在背后和“自己人”说别人的不是。 利众、真诚 利众、真诚是Google自成立之初就始终坚守的信条。美国“卡特里娜”飓风过后，一家与Google Map合作的卫星公司得到了最新的灾区卫星图片，并将图片发给Google。当时，Google Map小组的所有人都一致认为，尽快将这些卫星图片免费发布到网上，可以解决政府机构和灾区民众的燃眉之急。于是，他们彻夜不眠，用最快的速度更新了服务器中的地图信息，为救灾工作做出了贡献。 大部分公司上市都是经过投资银行，而投资银行又会把最多的股票分给大户低价买进，一两天后再卖出获得暴利，而小户可能就此被套牢。Google股票上市时，一位中国工程师沈修平提出：“为什么我们不用平等的方式来拍卖我们的股票，无论大户小户都可以用同样的价钱买入”。因此，Google自己做了一套软件来完成这个符合公司价值观和文化的上市流程，并赢得了广大股民的一致好评。这个例子很好地证明了，公司的每一件事都应该符合文化，只有这样，优秀的文化才会得以传承和延续。 激情、风趣 Google的员工都喜欢在愚人节那天开玩笑。有一年4月1日，有位Google员工甚至在网站上推出了Google牌的彩色汽水。这种提倡激情和趣味的企业文化体现在研发上面，就成了无数新点子、新创意的最好源泉。 在月球上建立研发中心是另一个愚人玩笑，还真有人上当了，要申请加入Google未来的月球研发中心。愚人节玩笑开多了，真的都被当成了假的。Gmail正好在4月1日发布，很多人以为这又是Google这帮聪明家伙的玩笑，而且，那时候并没有企业曾经免费提供如此大容量（1GB，现在超过了2GB）的邮箱。最初，这个消息甚至被不少内部员工也认为是假消息，可见愚人节玩笑后遗症之“流毒”深远矣。 一位刚加入Google的新员工，由于尚未租到房子，就在公司生活了一个月。他的生活远比汤姆汉克斯在机场的生活——电影《幸福终点站（The Terminal）》的故事情节——爽得多。一日三餐自然不用愁，到处都是可以睡觉的舒服的沙发，洗澡可以在十分高级的浴室完成，锻炼身体可以去折腾游泳机和跑步机，偶尔还可以在钢琴室里活动活动手腕，比在自己家里还舒服。据说他居然做到了一个月足不出户！ 人才：Google的“核心财富” Google自创立的那一天起，就把发现人才、吸引人才、培养人才、重视人才视做最为重要的管理理念，因为只有世界一流的人才才能为公司创造最大的价值。 在今天这样的信息时代里，人才的价值尤为重要。在工业时代，一个优秀技工和一个普通技工的效率差异可能是30%，但在信息时代，一个高级程序员和一个普通程序员的效率差异可能高达10倍以上。例如在Google，公司最顶尖的编程高手Jeff Dean曾发明过一种先进的方法，该方法可以让一个程序员在几分钟内完成以前需要一个团队做几个月的项目。他还发明了一种神奇的计算机语言，可以让程序员同时在上万台机器上用最短的时间完成极为复杂的计算任务。毫无疑问，这样的人才对公司来说是有非常特殊的意义的。公司必须善于追踪、挖掘、面试、聘请、评估、培养、重视和留住这样的人才，并总是把与人才相关的工作视为公司最重要的任务之一。 发现人才 重视人才的一个最直接的体现就是管理者要善于雇佣高水平的人才。在Google，每一个管理者最重要的一件事情就是招聘。有一次，记者问我们的首席执行官施密特，未来公司的成长瓶颈是什么？我想，许多其他公司的首席执行官在遇到类似问题时可能会回答市场前景、竞争、技术发展等等。但是，施密特当时不假思索就立即回答道：“我们能不能够继续找到世界一流的人才。” “招聘专员（recruiter）”这个职位在其他的公司可能只是人事部门里面的一个普通职位，一个公司最多也不过有几十个招聘专员。但是在Google，我们几千个员工中居然有500个招聘专员。之所以设置这么多专职负责招聘的岗位，那是因为我们深深理解，公司需要人才的岗位非常多，每一个应聘者又都要经过周密、严格的面试流程，所以，我们必须要有足够数量的招聘专员来完成这样繁重的工作。 我当初打算来到Google的时候，Google一知道我对Google有兴趣，一位资深副总裁就立即发电子邮件告诉我说：“随时打电话给我，一星期七天，一天24小时，任何时间都行！”为了我的方便，当我到Google去面谈的时候，他们特意选择了一个高尔夫球场作为谈话场所，而每一个参加面谈的资深员工也都不介意为了我一个人跑这么远的路——特别是公司的两位创始人拉里和谢尔盖，他们一个骑着自行车，一个踩着滑板来到了高尔夫球场。 Google公司坚信：一流的人雇佣一流的人，二流的人雇佣三流的人。So when a team hires its first second-rate guy, that's when it goes downhill.所以，公司强烈鼓励员工推荐自己认识的高手加入。公司有一半的人都是经过员工推荐进来的。最近，为了鼓励大家帮助一个新成立的部门招募人才，公司更是推出了诱人的奖励政策：只要你能成功地推荐十个人进入这个部门，就送你一辆车！ 总能发现并得到世界上最好的人才，Google似乎拥有一种不可抵御的“魔力”。我认识的许多朋友都先后来到了Google。他们中有资深的研究者和科学家，也有技术超群的工程师。而且，他们中的每个人一到了Google，就好像被充了电一样，浑身上下都充满了激情。这样一批优秀的人才，在这样优秀的公司里，怎么能不创造出一个又一个的优秀产品呢？ 在Google，几乎每个领域都有世界一流的大师，单单列举一下这些大师的名字就足以让人兴奋不已了：互联网之父Vint Cerf，世界上第一个网络爬虫（Crawler）的发明人Louis Monier，UNIX和C语言大师Rob Pike，Linux世界的第二号人物Andrew Morton，Apache组织的主席Greg Stein，Python创始人Guido Van Rossum，XML的发明人Adam Bosworth，中文搜索第一人简立峰……%%%选拔人才 众所周知，Google的面试机制既严格、精密，又充满了Google特色。 对于面试，Google采取了极为严谨的态度，使用统一的、经过精心设计的面试流程。例如，Google虽然在其他方面非常习惯放权给主管经理，但在招聘方面，每个团队的主管经理或招聘经理并没有最终的决策权。因为公司担心每个经理会因为自己希望尽快充实团队而放松了对招聘标准的把关。例如，很多人在招聘时会有类似于“在五个申请者中挑一个最好的”的想法。当应聘者都不能满足公司要求时，他就会不自觉地降低标准。因此，在Google，几乎所有应聘者都必须沿着统一的面试流程走完所有招聘步骤，然后被“招聘小组”客观评估。 例如，假设在一个星期内有一万人投简历，我们会先对简历做一次筛选，那些较优秀的会优先得到我们的电话面试（在中国也可能有笔试）。电话面试过关的，我们会安排面试。在面试中分数较高的人的资料可以被提交到每周的“招聘小组”讨论。最后，“招聘小组”挑出来的最优秀的人又被送到公司的“副总裁小组”再一次审查。这一万人里可能只有100人过关，我们这时会把录用通知发给这100个人，也许里面有90人接受。然后，接受录用的人就可以决定自己从哪一个月开始工作了。一直到这时候，各团队的主管经理才能来“争取”这些“应聘成功者”。同时，每一个刚加入Google的新员工也会根据个人爱好提出自己希望从事的工作列表，公司则会把那些最合适的人分配给最需要他们的部门或项目组。 在整个流程里，每个环节都是精心设计的。拿面试来说，除了极少数有特殊技能的应聘者之外，几乎所有的应聘者都会被不同部门的人面试。每一个面试官都可以看到此前的面试官提出过的问题（这样就不会有重复提问的情况），但是他看不到其他面试官的打分（这样才能保证每个人都足够客观）。每一个面试官在面试后要填写非常详细的反馈信息（通常在两页左右），然后还要为应聘者打分，分数从1.0到4.0不等（可以是小数，1.0表示绝对不能雇用，2.0表示反对雇用，3.0表示可以雇用，4.0则表示坚持必须雇用）。给分时，每一个面试官都必须十分谨慎，因为他的面试技能也会被打分和评估。如果一个面试官多次在别人都给低分的情况下给出高分（或者反之），他自己就会被扣分。如果他的判断经常和别人相反，那么他以后可能就不会有参加招聘的机会了！ 为了更好地了解应聘者的情况，公司内部还有一个自动化的软件，它会针对每一个应聘者的简历自动去询问那些可能认识该应聘者的公司员工（例如，这个软件如果发现了一位1995年在苹果公司从事多媒体工作的工程师，就会自动发邮件问我认不认识这个人，请我给出评价，因为该软件知道，当时我正在苹果公司负责多媒体工作）。 当所有反馈信息都搜集好了之后，“招聘小组”的工作就显得特别重要了。“招聘小组”和公司内的任何项目都没有直接关系，小组成员会审查关于应聘者的所有反馈信息，以及应聘者的学校、学历、工作经验、推荐信、推荐人等信息，然后对所有应聘者的分数进行排序。通常，如果一个应聘者得到了一个2.0，或两个2.5，或有非常明显的负面反馈（如难以沟通、诚信有问题等等），“招聘小组”就不会考虑他了。 这是一种独一无二的招聘模式，也许它并不是最有效率的，但它能够真正保证我们所雇用的都是一流的人。 Talent 在Google，因为我们招聘的大都是编程高手或算法大师，所以一般来说，我们并不鼓励他们成为经理，因为那样的话他们的编程、算法功力就无法充分发挥出来。所以，我们必须要让所有的工程师有一个和经理一样好的“双轨道”制度，也就是提供一套合理的晋升机制，让一个技术人才完全可以做到和副总裁、总监一样的级别。在不同的轨道上，Google为员工提供的机会是平等的，员工并不一定非要做管理工作才能促进个人事业的发展。在Google，一个最高级别的工程师可能比副总裁还要资深。Such a "dual-track" policy institutionally ensures the diversity of talent development paths, which is conducive to attracting and retaining talents. Google非常鼓励人才在公司内部流动和发展，员工根据自己的意愿选择或改变工作内容并不是一件很困难的事。例如，大约每六个月，工程师就可以考虑自己是不是要换一个新的项目，以便学习新知识或更好地发挥特长。因为Google的大多数项目都比较短，差不多每六个月，手头的工作就可以告一段落，同时Google在项目管理和知识共享方面做得也很出色，一般的项目都不会因为人员变动而受到影响。当然，这一切都是看个人兴趣决定。如果一位员工很热爱爱的项目，他可以长期地留在这个项目。如果一位员工只喜欢做最新、最有创意的工作，而他确实也很有创意，那么它就可以每六个月（甚至更频繁）开启新的项目。 Google总是选择那些最优秀的经理人来完成管理和招聘工作，同时尽量保持经理队伍在人数上的精简，以便让管理结构更为扁平，无论是经理还是普通员工，都可以在这样高效率的组织结构中充分发挥特长。也就是说，我们希望每个经理都能直接领导一个较大的团队（某些经理甚至直接管理一百多人的团队）。因为这些经理人都非常优秀，这种扁平的管理方式实施起来并不很困难。当然，这样做的一个问题是，经理可能不会详细掌握每个员工的工作情况。于是，我们又建立了员工间相互评审的制度，即每个员工在接受考评的同时，也要对他身边同事的工作表现做出评价。相互评审中的许多内容都是公开的，评审者和被评审者都能看到。经理可以收集所有自我评估和相互评审的反馈信息，然后给出最终的评审意见。 大多数公司都会根据员工业绩和末位淘汰的原则，定期解雇一些业绩较差的员工。但Google几乎从不主动解雇员工。因为我们相信，我们在招聘时采用的面试标准已经非常高了，只要能够通过面试，进入Google的员工就一定是出色的人才，没有必要施行末位淘汰制度。 留住人才 在Google，人才流失率不到1%，相比之下，美国硅谷的平均人才流失率约为20%。这时因为Google非常重视发挥人才的特长，为他们提供优越的工作环境，以便留住这些优秀的人才。 很多人可能认为，留住人才的关键是待遇。Indeed, everyone should be treated appropriately, but for a software practitioner, what he values ​​more is the working environment.与提高员工待遇相比，Google更注重创建一个能够吸引人、留住人的优越环境，这包括： * 充分的资源支持，让每个人都没有后顾之忧。例如，Google中的每个人都可以得到他们需要的性能最好的电脑，以及两个24英寸大屏幕的液晶显示器。 * 吸引人的科研工作，让每个人都热爱自己的工作。每个工程师都可以根据自己的意愿在适当的时候变换项目，以便更好地发挥特长。 * Leaders understand and support employees' research, so that everyone can follow the general direction of the company while still having enough space and freedom to develop their talents and pursue their dreams. * 创新实践者的乐园 * 在Google，喜欢创新的人可以通过自己的智慧，创造出最先进的、最有价值的网络服务产品，真正把互联网变成造福世界的平台。 * 在Google，一切都可以自己创造，一切也都需要自己动手（因为工程师最喜欢的就是自己动手，创造奇迹），甚至连Google使用服务器都是自己组装和定制的（与购买现成的服务器，这可以节省大量资金）。 * 自由创新，改变世界 * Google将世界看做自己的实验室，并尽力在最短的开发周期内推出最好的产品，“今天就改变世界”是所有Google工程师的坚定信念。 * Google的创新思维和优越环境可以让员工充分发挥才华和潜力，做出最大的贡献。 * 如果想创业，员工甚至不必离开公司。因为Google内部每年都会有几个项目得到Founder's Award，奖金大约美元100万美元，和出去创业所能得到的也差不多，而且没有创业的风险，周围还可以有那么多杰出的人一同工作。 * 为员工提供学习、培训的机会 * 正如上文所列举的那样，Google靠自己的影响力吸引了大批世界一流的大师，并借助他们的智慧、经验培训和影响新员工。事实上，这些大师的存在就已经意味着出最好的学习和工作环境了。 * Google为员工提供了丰富的培训课程，包括内部课程，科学或人文的讲座，也包括支持员工（包括中国员工）到斯坦福大学读硕士学位等等。 * 新颖的环境和福利 * 优美的办公园区，新颖的办公室装饰，诱人的美食，漂亮的健身房，舒适的按摩服务，各式各样的游戏和娱乐装置，通过媒体的宣传，这些典型的Google文化对大多数人来说已经不再陌生了。在这样优越的环境中工作，员工的工作热情怎能不高涨，工作业绩怎能不出色呢？ * 百步之内必有食物！这是Google内部一条规矩。一位工程师觉得并没有做到这一点，就量了一下他与最近的食物之间的距离，果真发现超过了100英尺，距离是120英尺。于是他幽默地要求后勤部门改进自己的服务，对方的回答十分经典：你的头顶就是餐厅，距离你绝对没有超过100百步！ 企业员工是Google的主人，工程师是Google的灵魂。在Google工作的员工可以享受到最好的工作环境和福利，但更为重要的是，员工可以非常容易地找到发挥特长的机会，以便把自己的技术理想变成千万人使用的热点产品——对那些渴望在轻松、愉悦的环境中实现自身价值的技术天才来说，这样的人文环境该有多么强烈的诱惑呀！ conclusion Google公司是高举创新旗帜，走在互联网时代最前沿的创新型企业。Google公司的创新不仅仅是产品和技术的创新，更是文化和管理的创新。正是因为有了平等互重、创新实干、透明客观、利众、真诚、激情风趣、以人为本等优秀的企业文化，有了全力发现、选拔、重视和吸引世界一流人才的坚定理念，Google才实践出了一条符合网络时代特点的创新之路。 希望Google这种集研发一体、个人自由、平等参与、实时高效、用户优先和使命清晰等特点为一身的创新模式，以及符合互联网时代特点的企业文化和人才理念能够为更多的IT企业所借鉴，希望大家都能根据企业自身的特点，创造出符合实际情况的、创新的管理方式和企业文化，并在此基础上研发出最适合企业发展需要，也最能满足未来时代要求的新技术和新产品。